Outdoor 3D Environment Reconstruction based on Multi-sensor Fusion for Remote Control

Abstract

Environmental awareness is the necessary ability for unmanned driving. In some specific scenarios, remotely controlled intelligent vehicles can perform tasks more efficiently. The remote operators need to make quick decisions based on the obtained environmental information. The result of 3D outdoor reconstruction can provide the operators with three-dimensional environmental information. In this paper, a 3D reconstruction framework based on multiple sensors is designed for intelligent vehicle to be aware of driving environment. Through the joint calibration of the camera and the LiDAR (Light Detection And Ranging), the image and the 3D point clouds are combined to obtain the fused colorful point clouds. The first-fused result is combined with the real-time vehicle pose to realize the three-dimensional reconstruction of the intelligent vehicle outdoor driving environment by splicing the point clouds. The final 3D reconstruction effect is displayed using the computer graphics technology in real time. The proposed solution was carried out on the built experiment platform to verify the visualization performance.